Until now, the neurodevelopmenral function and role of the human ROBO1 gene have remained unclear, although it is suggested to predispose to dyslexia (Nopola-Hemmi et al., 2001; Fisher et al., 2002; Hannula-Jouppi et al., 2005) and shows linkage to autism (Anitha et al., 2008) and phonological processing (Bates et al., 2011). In addition, a region around the ROBO1 gene has been associated with a speech sound disorder, an SLI-variant (Stein et al., 2004). The underlying causes of all these disorders—each of them relating at some level to language and auditory processing—are unknown.
Our results of impaired binaural interaction in ROBO1-deficient subjects revealed for the first time the significance of the ROBO1 gene in sensory processing and brain function. The correlation between the binaural suppression and the expression level of ROBO1 suggests that, for normal binaural processing, adequate expression level is needed. Studies in fruit flies and mice have demonstrated the critical roles of robo and Robo1, orthologs of human ROBO1, for axonal midline crossing during development (Kidd et al., 1998a; Kidd et al., 1998b; Brose et al., 1999; Andrews et al., 2006).
Because binaural interaction requires axonal midline crossing—otherwise the inputs of the LE and the RE would not converge—and because normal auditory pathways show extensive crossing in the brain stem, the found impaired binaural interaction suggests defective crossing of auditory pathways. During embryonic development of the rat nervous system, Robo1 mRNA has been found in many places around the auditory system (Marillat et al., 2002): in CNs and ICs, both sending crossing axons to the
52
opposite side (Moore, 1991), and in medial part of the dorsal thalamus and in lateral cortices. Above the tectum, the auditory pathways remain totally or near-totally ipsilateral, and therefore the defective crossing may take place already in the CNs or in the ICs. The ICs are one of the most important structures for binaural processing (Moore, 1991).
In our ROBO1-deficient subjects, dyslexia co-segregated with the ROBO1 gene defect in a dominant manner. Dyslexia is a complex neurodevelopmental disorder with strong genetic background. Until now, six genes, DYX1C1, DCDC2, KIAA0319, C2Orf3, MRPL19, and ROBO1 have been associated to dyslexia susceptibility (for a review, see Kere, 2011; Peterson and Pennington, 2012), and linkages to several additional genetic regions are weakly established (for a review, see Scerri and Schulte-Körne, 2010;
Peterson and Pennington, 2012). Animal studies have clarified the roles of the DYX1C1, DCDC2, KIAA0319, and ROBO1 in neuronal migration and axon guidance during embryonic development (for a review, see Galaburda et al., 2006; Scerri and Schulte-Körne, 2010; Kere, 2011), but they have not managed to bring knowledge about their role in sensory or any other neural processing.
Previous neuroimaging studies in healthy humans have associated DYX1C1, DCDC2 and KIAA0319, but not ROBO1, to brain morphology. Specific polymorphisms of DYX1C1, DCDC2, and KIAA0319 were associated with white matter density in left temporo-parietal area (Darki et al., 2012) and DCDC2 genotypes were associated with grey matter volumes in superior, middle and inferior temporal gyri in the LH, and also with fusiform, hippocampal/para-hippocampal, inferior and middle frontal, and inferior occipito-temporal gyri in the LH (Meda et al., 2008). In an fMRI study, genetic variants of the KIAA0319 locus were associated both with activation and hemispheric asymmetry of activation of posterior STS during reading (Pinel et al., 2012). Before the Study I, only one study of dyslexic subjects has examined the connections between the dyslexia candidate genes and neurophysiology: in German children, auditory late mismatch negativity (MMN), 400–600 ms after syllable onset, was associated with four specific variants in a genetic region containing KIAA0319 and DCDC2 (Czamara et al., 2011). Late MMN, on the other hand, has been suggested to be related to letter-speech integration (Froyen et al., 2009), and in dyslexic individuals late MMN amplitudes to speech stimuli have been small (Schulte-Körne et al., 2001).
53
Altogether, earlier data, and especially human data, about the functions of dyslexia susceptibility genes are sparse, and our result of strongly suppressed binaural interaction implying defective crossing of auditory pathways provides novel functional and anatomical knowledge about the interaction between the human sensory nervous system and dyslexia susceptibility gene. It remains unclear whether the found impairments in binaural processing parallel, contribute, or even cause some symptoms of dyslexia.
Interestingly, in dyslexic subjects, enhancement of brainstem responses to repetitive speech sounds is abnormally weak and this enhancements correlates with speech understanding in noise (Chandrasekaran et al., 2009). On the other hand, speech perception is better with two ears than one ear, most probably because of binaural interaction (Cherry, 1953).
During the last decades, a variety of phonological, visual, auditory, attention, and tactile deficits have been found in dyslexic subjects and many different hypotheses of dyslexia have been suggested. Many of the found deficits may result from the defective function of certain dyslexia susceptibility genes, but are not necessarily causally related to the development of dyslexia. Interestingly, all studies of dyslexia candidate genes thus far have shown linkage to the temporal lobes. However, they have not managed to reveal the underlying brain functions and e.g. the role of auditory processing deficits in development of dyslexia remain unclear, although their possible causal role has been again highlighted (Galaburda et al., 2006; Goswami, 2011).
To show whether the impaired binaural interaction associates only with ROBO1 gene expression or also with some other dyslexia susceptibility genes, comparisons between dyslexics with different genetic backgrounds might be informative. In Study I, no correlation analyses between the binaural interaction and ROBO1 gene expression in healthy controls were carried out, although the controls showed wide individual variability in binaural interaction levels. Therefore, it remains unclear whether the ROBO1 expression also correlates with binaural interaction in the general population. In future studies, the relationships between binaural interaction, different phonetic deficits, speech perception in background noise, and also sound localization ability may be clarified. In addition, studies of binaural interaction with speech sounds might provide further information about binaural processing. However, revealing—or proving to be
54
false—the possible link between the auditory deficit and later developmental defects is challenging. The central auditory system can undergo remarkable plastic changes and compensate for many disturbances; e.g. commissural pathways and association axons, the important connections between different auditory areas and thus to the perception of complex auditory stimuli, mature strongly during childhood and early adolescence (Moore, 2002), and reading ability, as any acquired skill, also modifies the brain (Dehaene et al., 2010). Thus, the original underlying defects may disappear or fade during development and may not be measurable in adults, whereas the other non-causal symptoms may even strengthen (Galaburda et al., 2006).
55